Google Discover has become a lifeline for many publishers as classic blue-link search sends fewer visits. At the same time, Google is injecting more “in-feed consumption” through AI-powered summaries, which can satisfy curiosity without a click and weaken the discovery-to-session funnel that publishers rely on for ads, subscriptions, and loyalty.
Protecting Discover traffic from AI overviews isn’t about fighting AI; it’s about shaping how your content is previewed, cited, and differentiated when Google summarizes. The goal is to keep your brand and value proposition visible, preserve click incentives, and reduce the risk that your best work becomes a “free answer” inside a card.
1) Why AI Overviews change the economics of Discover
AI Overviews in Google Search have been associated with significant click-through-rate declines. Ahrefs reported (Feb 6, 2026, based on December 2025 data across 300,000 keywords) that when an AI Overview appears, CTR to the #1 result can drop by around 58%, and earlier analysis summarized by eMarketer (Apr 18, 2025) found a 34.5% lower average CTR correlated with AI Overviews.
Publishers connect those SERP dynamics to downstream distribution channels like Discover because overall Google referrals are already under pressure. Digital Content Next data summarized by eMarketer (Aug 15, 2025) linked AI Overviews to publisher referral declines up to roughly 25% (with many outlets seeing smaller but meaningful drops), reinforcing why every remaining surface, including Discover, must be defended carefully.
Google’s own product direction for Discover (Sep 17, 2025) frames it as a “jumping-off point for exploring the web,” with more formats and ways to follow creators. That positioning creates opportunity, but it also signals a future where more reading may happen “inside the feed” unless publishers actively protect their incentives to click.
2) AI-generated summaries are now inside Discover (and they can absorb the click)
Google has rolled out AI-generated summaries inside Discover in the US, where cards can show an AI-written overview with multiple-source icons. Tapping can reveal cited articles, and Google includes a warning that summaries are AI-generated and may be inaccurate, yet the interaction still encourages consumption in-feed before a user commits to visiting any publisher site.
Industry reaction has been cautious: TechCrunch reported publishers fear Discover AI summaries could further reduce referrals by satisfying intent without a click. That concern mirrors broader reporting about AI search being “devastating” to publisher traffic, including TechCrunch’s June 10, 2025 summary of WSJ/Similarweb-referenced declines.
In practice, Discover is a visual, fast-scrolling environment. If the user can get “good enough” understanding from a summary plus a glance at cited sources, the marginal benefit of tapping through to a full article shrinks, especially for topics that are informational, evergreen, or easily summarized.
3) Use Google’s preview controls to reduce “answer theft” effects
Google Search Central’s guidance on “AI features and your website” explains that site owners can use preview controls to influence how content appears in AI-powered search experiences. Those controls include directives like nosnippet, max-snippet, max-image-preview, and noindex, giving publishers levers to decide what can be quoted or previewed.
For protecting Discover traffic, the practical idea is to limit how much of your article can be “consumed for free” in summaries and previews. A max-snippet cap can reduce how much text Google may show, while nosnippet prevents text snippets entirely, potentially increasing the need to click, at the cost of less context in the feed.
These directives are not one-size-fits-all. Apply them surgically: protect premium, exclusive, or highly summarizable content (where the overview can replace the visit), while leaving snippets available for content where the snippet functions as a teaser and improves qualified clicks.
4) Protect Discover CTR by controlling image previews (max-image-preview applies to Discover)
Google’s robots meta tag specifications explicitly state that max-image-preview applies to “all forms of search results (such as Google web search, Google Images, Discover, Assistant).” This matters because Discover is highly visual, and large thumbnails are a major driver of card-level engagement and CTR.
Counterintuitively, limiting large thumbnails can sometimes protect clicks when AI summaries are present. If a Discover card becomes too “complete” (big image + AI summary + enough context), the user may feel no need to visit the page. Using max-image-preview:standard or even none is a direct lever to reduce overly rich previews in surfaces including Discover.
That said, images also attract attention; shrinking them can lower reach if competitors keep large previews. Treat max-image-preview as a testing variable: run controlled experiments by section, template, or content type, and measure not just CTR but downstream quality (time on site, recirculation, conversions).
5) Optimize for being the cited source, without giving away the whole story
If Discover AI summaries cite multiple sources, your goal shifts from “just rank” to “be selected and clicked.” Research by Aral, Li, and Zuo (arXiv, Feb 13, 2026) reports expanding exposure of AI Overviews and finds AI search surfaces fewer long-tail sources and changes the mix of variety and credibility. That implies larger brands may be favored unless smaller publishers make their authority and uniqueness unmistakable.
To increase the odds that the summary needs you, build “non-substitutable” value into the article: original reporting, proprietary data, firsthand quotes, unique visuals, tools, and clear attribution. The more your piece contains information that cannot be confidently paraphrased without losing credibility, the more likely users are to click through for details and verification.
Also write with citation in mind. Clear ings, concise definitional passages, and well-labeled original findings make it easier for systems to understand what is uniquely yours, while preview controls help ensure those passages don’t become a complete replacement for a visit.
6) Prepare for volatility: UI changes, safety issues, and rapid product shifts
Google has signaled it is adjusting AI experiences to make source links more obvious. The Verge reported planned UI changes such as more visible link affordances and richer source previews (e.g., hover/desktop pop-ups with descriptions/images), which could partially mitigate click loss by making attribution and navigation easier.
At the same time, safety and quality controversies can accelerate rollout changes. Wired reported AI Overviews being exploited for scams (like fake phone numbers), and The Guardian highlighted concerns about health disclaimers being visually de-emphasized. These incidents increase uncertainty: features may expand, retract, or change presentation quickly, reshaping referral patterns with little notice.
Operationally, treat AI summaries in Discover as a moving target. Maintain a monitoring routine (traffic segmentation for Discover vs Search, snippet/preview changes, template audits) and keep “rapid response” options ready, like tightening snippet caps on vulnerable sections, or loosening previews where visibility is dropping too far.
7) A practical playbook to defend Discover sessions
Start by mapping which content is most at risk of being fully summarized: explainers, “what is,” “how to,” and quick-answer posts. Those are prime candidates for max-snippet limits, selective nosnippet, and more deliberate lead-writing that teases the payoff rather than giving it away in the first lines.
Next, use Discover-specific creative discipline. If AI summaries appear on a card, your line and image must communicate why the full article is worth the tap. Emphasize what the summary cannot deliver: original documents, interactive elements, deep comparison tables, expert interviews, or local/on-the-ground context.
Finally, diversify your “post-click” value so clicks translate into durable audience. Improve recirculation, newsletter capture, and membership prompts on Discover landing pages, because in an environment where AI Overviews correlate with lower CTR and DCN reports notable referral declines, each session must be more valuable than it used to be.
Discover remains a critical distribution channel, but AI-generated summaries inside the feed raise the bar: publishers must actively choose how much they let Google preview, quote, or visually package their work. Google’s own documentation confirms you have tools, preview controls and image preview directives, to shape that outcome.
The most resilient strategy combines technical controls (max-snippet, nosnippet, max-image-preview) with editorial differentiation and ongoing measurement. As Google iterates on AI UI and responds to safety concerns, the publishers who treat Discover optimization as a living system, not a one-time SEO task, will be best positioned to protect Discover traffic from AI overviews.
